Mr. Ferrier has shown that "the senses are only _contingent conditions_ of knowledge; in other words, it is possible that intelligences different from the human (supposing that there are such) should apprehend things under other laws, or in other ways, than those of seeing, hearing, touching, tasting, and smelling; or more shortly, _our_ senses are not laws of cognition or modes of apprehension which are binding on intelligence necessarily and universally."--"A contingent law of knowledge" is defined as "one which, although complied with in certain cases in the attainment of knowledge, is not enforced by reason as a condition which _must_ be complied with wherever knowledge is to take place. Knowledge is thus possible under other conditions than the contingent laws to which certain intelligences may be subject; in other words, there is no contradiction in affirming that an intelligent being may have knowledge of some kind or other without having such senses as we have."[285] The application of analogy as a principle of judgment is subject to certain well-known limitations, which cannot be disregarded without serious risk of error. They are well stated by Dr. Hampden: "There are two requisites in order to every analogical argument:--1. That the two, or several particulars concerned in the argument should be known to agree in some one point; for otherwise they could not be referable to any one class, and there would consequently be no basis to the subsequent inference drawn in the conclusion. 2. That the conclusion must be modified by a reference to the circumstances of the particular _to_ which we argue. For herein consists _the essential distinction between an analogical and an inductive argument_. Since, in an inductive argument, we draw a general conclusion, we have no concern with the circumstantial peculiarity of individual instances, but simply with their abstract agreement. Whereas, on the contrary, in an analogical argument, we draw a particular conclusion, we must enter into a consideration of the circumstantial peculiarity of the individual instance, in order to exhibit the conclusion in that particular form which we would infer. Whence it follows, that whilst by induction we obtain absolute conclusions, by analogy we can only arrive at relative conclusions, or such as depend for their absolute and entire validity on the coincidence of _all_ the circumstances of the particular inferred with those of the particular from which the inference is drawn." Again: "The circumstances _to_ which we reason may be considered of threefold character. They are either known or unknown. If they are known, they are either (1.) Such as we have no reason to think different, in any respect from those under which our observations have been made; or (2.) Such as differ in certain _known_ respects from these last. (3.) They are unknown, where we reason concerning truths of which, from the state of